LANGUAGE IN INDIA

Strength for Today and Bright Hope for Tomorrow

Volume 1: 8 December 2001
Editor: M. S. Thirumalai, Ph.D.
Associate Editor: B. Mallikarjun, Ph.D.

  • HOME PAGE

  • DECEMBER 2001 ISSUE

  • E-mail your articles and book-length reports to thirumalai@bethfel.org or send it by regular mail to:
    M. S. Thirumalai
    6820 Auto Club Road #320
    Bloomington, MN 55438 USA
  • Contributors from South Asia may send their articles to
    Dr. B. Mallikarjun,
    mallik_ciil@hotmail.com.

  • Your articles and booklength reports should be written following the MLA, LSA, or IJDL Stylesheet.

  • The Editorial Board has the right to accept, reject, or suggest modifications to the articles submitted for publication, and to make suitable stylistic adjustments. High quality, academic integrity, ethics and morals are expected from the authors and discussants.

LEXICAL FORMATIVES AND WORD FORMATION RULES IN TAMIL

N. Ramaswami, Ph.D.

1. INTRODUCTION

One of the main functions of the theory of morphology is to define the notion 'possible word in a language.' While defining the possible word in a language the generative morphology has accepted lexical items as base on which Word Formation Rules (WFR's) operate.

There are a few proposals regarding the basic units of the lexicon. According to Halle (1973) morphemes are the basic units of the lexicon whereas according to Aronoff (1976) words are the basic units of the lexicon. In other words, morphology is morpheme-based and word-based according to Halle and Aronoff respectively. However, it seems that the Word-based Hypothesis is subject to typological differences (Hoestra et al 1979:20 as stated in Scalise 1984:74). The position to consider both stems and words as the units of the lexicon may be taken to be the compromising stand acceptable to many. The stems maybe considered as a special sub-part of the dictionary.

This paper is divided into three parts: the first part tries to establish the basic units of Tamil lexicon; the second part tries to put forth a few restrictions on the attachment of suffixes and words to the base in order to derive e grammatical words and to exclude ungrammatical combinations; and the third part tries to propose some Word Formation Rules for Tamil.

2. THE BASE

The lexical component of a grammar has a list of words and Word Formation Rules. The Word Formation Rules may be divided into Derivational Rules, Compounding Rules, and Inflectional Rules. There are also Readjustment Rules, which operate on the output of Word Formation Rules to derive the acceptable phonological shape of the words. The Readjustment Rules may be divided into two: Truncation Rules and Allomorphy Rules. The model of Dictionary proposed here for Tamil will have words and stems. Both the words and stems are bracketed and labeled. Stems do not have internal boundaries but the words have internal boundaries. Words will be labeled syntactic category and will appear on the surface freely since they are free forms. Stems will also be marked with syntactic category but they cannot appear on the surface since they are not free forms.

Words may be divided into simple, derived, and compound. Simple words may be any free form without having any derivation. The derived word may be another simple word or a compound. So the proposed Dictionary may have simple words, derived simple words and compounds.

There are no prefixes in Tamil Words are free forms but stems and suffixes are not free forms. Suffixes1 will be attached to words and stems. Stems will be attached to the words as in mara + niLhal 'shade of the tree,' paNa + peTTi 'cash box,' and words may be attached to words to form compounds as in maalai # neeram 'evening time,' cuTar # viLakku 'bright lamp.'

3.1. SIMPLE WORDS

Simple words will not be marked with internal boundaries but will be labeled with syntactic categories (Scalise 1984:78). The representation of simple words, in Tamil, may be as follows:

[tamiLh]N, [col]N, [paar]V, [periya]adj, etc.

3.2. STEMS

Stems2 are bound forms, but words are not.

As simple words, stems also will be labeled with syntactic categories but there will not be internal boundary. ooTu 'to run,' aaLu 'to rule,' eLhutu 'to write,' maaTu 'cattle,' kiNaru 'well,' etc. are simple words since there are no internal brackets and they can stand independently giving full meaning. When these words occur in certain constructions, their phonological shape changes as shown below:

1. avan kaarai ooTTinaan
'He drove the car.'
2. raav naaTTai aaL-kir-aar
'Rao is ruling the country.'
3. neeru naaTTai aaNTaar
'Nehru ruled the country.'
4. inta eLhuttu teriyavillai
'This letter is not visible.'
5. atu oru maaTTu vaNTi
'That is a bullock-cart.'
6. anta kiNaru niir nanRaakaillai
'The water of the well is not good.'

3. 3. SUFFIXES

The suffixes may include both prefixes/infixes and suffixes.

Since suffixes are attached to other formatives, they will not be bracketed as autonomous formatives. They will be marked with boundary symbols (Scalise 1984:79). They will not be labeled with only one syntactic category since they show a relation between two syntactic categories. Since Tamil does not have prefixes, this will be applicable here only to suffixes. The attachments of suffixes in the formatives may be seen here.

[paTi]V + ppu]suf}, [paar]V + vai]suf}, [neer] + mai]suf}

4. 1. RESTRICTIONS

WFR's are of the lexicon and they operate within the lexicon. They operate on a set of words that are the bases. In order to derive grammatical words and avoid deriving ungrammatical words, WFR's must have some kind of restrictions. This may help to determine the sort of information WFR's can access to, and the sort of operations they can perform.

4.2. SYNTACTIC RESTRICTIONS

The noun forming suffixes such as [-kaaran] and [-kaari] are attached to nouns. These bases cannot be attached to adjectives or verbs. So, the base has to have the syntactic information particularly the sub-categorization features in order to put the restriction in the operation of WFR's. The base may be represented as:

1. { [veelai]N + [kaaran]suf}N
2. {[veelai]N + [kaari]suf}N
3. { [kolai]N + [kaaran]suf}N
as against
4. * [karuppu]adj + [kaaran]suf
5. * [paaTu]adj + [kaaran]suf
6. * [paTi]adj + [kaaran]suf

However, [veLLai] + [kaaran] is possible but not
*[karuppu]adj + [kaaran]suf. It seems that [veLLai kaaran] is like an attributive construction as in [veeTTai + kaaran], [tooTTam] + [kaaran].

[kuTi kaaran] is also possible in which the formative [kuTi] is used as noun rather than a verb though the formative can function both as noun and verb.

The suffixes [-tal] and [-ttal] are attached to verbs but not nouns. Further, the suffix [-tal] is attached to intransitive verbs and the suffix [-ttal] is attached to transitive verbs. So the base must have the sub-categorization features: the suffix [-tal] may be attached to those verbs that are marked [-transitive] and the suffix [-ttal] may be attached to those verbs that are marked [+ transitive]. This may be seen in the following examples.

7. [piri]V-intr + tal] 'separating' (intr.)
8. [piri]V-tr + ttal] 'separating' (tr.)
9. [kalai]V-intr + tal] 'disordering' (intr.)
10. [kalai]V-tr + ttal] 'disordering' (tr.)
11. [eri]V-intr + tal] 'burning' (intr.)
12. [eri]V-tr + ttal] 'burning' (tr.)
13. [vaLar]V-intr + tal] 'growing' (intr.)
14. [vaLar]V-tr + ttal] 'growing' (tr.)

Nominalizing suffix [-tanam] is attached mainly to adjectives to derive nouns. This may be seen in the following examples:

15. [periya]adj + tanam]suf 'generosity'
16. [nalla]adj + tanam]suf 'good quality'
17. [kaLLaa]adj + tanam]suf 'stealthiness'
18. [kanca]adj + tanam]suf 'stinginess'
19. * [kaLLan]N + tanam]suf
20. * [paaTu]V + tanam]suf
21. * [kancan]N * tanam]suf
22. [ooTu]V + tanam]suf

However, [-tanam] is attached to the derived noun [putti + caali] 'intelligent' which will be used as noun also but in [putti + caali # payyan] 'intelligent boy' [putticaali] is an adjective. It is also attached to the nouns like kuRumpu 'mischief,' pookkiri 'rowdy' etc. as in kooLhai + tanam 'cowardice,' kaaTu + tanam 'rudeness.'

4. 3. SEMANTIC RESTRICTIONS

Since this area of semantic restrictions on the base has not been developed, it is difficult to go into it. See the following derivations of compound nouns.

1. { [kaar]adj # [meekam]N }N 'black cloud'
2. { [kaar]adj # [iruL]N }N 'black night'
3. { [kaar]adj } [kaalam]N }N 'rainy season'

The formative [kaar] refers to 'black' in the first two examples and in the third example it refers to 'rainy.'

The meanings of the two words may be accounted as this:

4.
(black) kaarmeekam 'black cloud'
kaar↑↓
(rainy) kaarkaalam 'rainy season'

This kind of situation arises when one of the constituents of the compound is polysemous. That is, the constituent behaves differently when they have more than one meaning.

4. 4. Phonological Restrictions3

1. Words in written Tamil do not have initial consonant clusters. Even if the borrowed words have initial consonant clusters they will be written without clusters by inserting a vowel in between the cluster as in kraaman à kiraamam 'village.' This restriction may be stated as

1. * # CC -

This condition states that word initial consonant clusters are not acceptable.

All words in written Tamil either end in a vowel or in a sonorant. No word ends in an obstruent. This restriction may be stated as

2. * # - [-son] #

This condition states that word final non-sonorant consonant is not acceptable.

However, a few borrowed words like bas 'bus' and TikeT 'ticket' end in obstruents.

2. Tamil has four past tense markers. They are [-t-], [-tt-], [-nt-] and [-in-]. The past tense suffix [-in-] is added to the verb base having the phonological shape ( C ) VVC ( C ). This may be seen in the following examples:

1. aaT - in- aaL 'danced (she)'
2. ooT - in - aan 'ran (he)'
3. paaT - in - aaL 'sang (she)'
4. teeT - in - aan 'searched (he)'
5. tuuNT - in - aan 'urged, motivated (he)'
6. veeNT - in - aan 'requested (he)'
As against
7. * oTi - in - aan
8. * kuTi - in - aan
9. * koTu - in - aan
10. * cey - in - aan
11. * paTi - in - aan

The examples given from 7 to 11 are not acceptasble since the phonological shape of the base of these words is not (C) VVC (C).

5. 1. WORD FORMATION RULES

1. It has been stated earlier that the base may be a stem or a word to which either a suffix or a word may be attached to derive a new word. WFR may specify the suffix or word to be attached to the base. In the process of attaching suffixes or words to these basic lexical formatives, either one suffix may be attached to the base, or two suffixes may be attached, or another lexical formatives may be attached, or another lexical formative followed by a suffix may be attached to the base. In Inflectional Morphology, many suffixes may be attached to the base. Some of the suffixation processes may be seen in the following examples:

1. x4 + suf
1. [ [eLhut-]V + tulsuf]N 'script'
2. [ [paaT-]V + tulsuf]N 'song'
3. [ [ooT-]V + tulsuf]N 'cause to drive'

2. x + suf + suf

1. { [putti]N + caali]suf + tanam]suf }N 'intelligence, quality of being intelligent'

3. x # x

1. { [ kooTTai ]N # kaTTu ]V}V 'nurture ambition'
2. { [eri]V + naTcattiram]N }N 'meteor'

4. x + suf # x

1. { [putti]N + caali]suf # [payyan]N }N 'intelligent boy'
2. { [aLhak-]N + aana]suf # [peN]N }N 'beautiful girl'
3. { [kaaT-]N # Tu]suf # [vaLhi]N }N 'forest route'
4. { [aLhai-]N + ppu]suf ## [itaLh]N }N 'greeting card'

5. x # x # x

1. { [kaar]adj # [kaala}N # [meekam]N }N 'The cloud of the rainy season'
2. { [maarkaLhi]N # [maata}N # [pani]N }N 'The mist of Markazhi month'

5. 2. CREATING NEW WORDS

Any new word created by a WFR must be the member of a major lexical category. The WFR determines the category of the output of the rule. In Tamil, the grammatical category may change or may not change after the operation of WFR. The following is the list of inputs and outputs of different kinds of WFR's in the derivation of simple words in Tamil.

1. Noun → Noun

1. [ [veelai ]N + kaaran ]suf ]N 'servant'
2. [ [toLhil ]N + aaLi ]suf ]N 'laborer'

2. Verb → Noun

1. [ [paTi ]V + ppu ]suf ]N 'education'
2. [ [eLhut-]V + tu ]suf ]N 'letter'
3. [ [keeL ]V + vi ]suf] N 'question'

3. Adjective → Noun

1. [ [nalla ]adj + tanam ]suf ]N 'good quality'
2. [ [periya ]adj + tu ]suf ]N 'big'

4. Noun → Verb

1. [ [uyir ]N + ppi ]suf ]V 'to give life'

5. Adjective → Verb

1. [ [veLLai ]adj + aakku ]suf ]V 'to make (something) white'
2. [ [karuppu ]adj + aakku ]suf ]V 'to make (something) black'

6. Verb → Verb

1. [ [cey ]V + vi ]suf ]V 'cause to do'
2. [ [naTa ]V -tt]suf - u] ]V 'cause to walk'
3. [ [viTu ]V + vi ]suf ]V 'to liberate'

7. Noun → Adjective

1. [ [uyaram ]N + aana ]suf ]adj 'high'
2. [ [aLhaku ]N + aana ]suf ]adj 'beautiful'
3. [ [neermai ]N + aana ]suf ]adj 'honest'

8. Verb → Adjective 0

9. Adjective → Adjective 0

There are no suffixes, in Tamil, to derive Adjectives from verbs and adjectives from Adjectives.

10. Verb → Adverb

1. [ [ cey]V + tu ]suf ]adv 'having done'
2. [ [ eLhut-]V + i ]suf ]adv 'having written'
3. [ [ paTi]V + ttu ]suf ]adv 'having read'

6. COMPOUND

1. Compounding rule is also WFR. Any compound may be represented as follows:

1. { # [ ]A B [ ]C + }D

2. 'A' and 'C' are the lexical formatives. 'D' is the grammatical category of the compound and 'B' is the internal boundary. There are two conditions regarding compounding. They are the Variable R condition, and the 'IS A' condition. The variable R condition states that the new compound has a range of possible meanings and the 'IS A' condition states about the combination of two constituents (Scalise 1984: 91-92). It seems that the first condition does not have much scope in Tamil. However, this may be seen later. The second condition may be stated as this:

2. { # [ ]A B [ ]C # }D

This condition states that there are two constituents 'A' and 'C' and the resulting compound is d in which the grammatical category of 'D' is 'C'. That is, 'D' IS A C. From the syntactic point of view the grammatical category of the compound 'D' is the same as that of the constituent C. This may be seen in the following formulas.

No. A B C
1 N N N
2 N V V
3 V N N
4 Adj N N
5 Adv V V

These conditions state that if the compounds are by the combination of the constituents like N + N, N + V, V + N, Adv + V, and Adj + N then the grammatical categories of the compounds will be N, V, N, V and N which are as that of the second constituents. This may be seen in the following examples. These conditions help to avoid writing different rules for each type and help to capture the generalization regarding compound formation and the grammatical categories of the constituents.

1. { [ oLi ]N # [viLakku ]N # }N 'bright lamp'
2. { [karuNai] N # [ manu] N # }N 'mercy petition'
3. { [Kalvi] N # [kuuTam ]N # }N 'educational institution'
4. { [ kural ]N # [ eLhuppu ]V # }V 'raise one's voice'
5. { [paNi ]N # [puri ]V # }V '(perform) work'
6. { [paNi] N # [aaRRu ]V # }V '(perform) work'
7. { [paLhi ]N # [tiir] V # } V 'take revenge'
8. { [ paRai ]N # [caaRRu] V # }V 'proclaim, make widely known'
9. { [koolam ]N # [pooTu] V # }V 'draw decorative designs on'
10. { [eri ]V # [ naTcattiram] N #} N 'meteor'
11. { [ eLhutu] V # [kool] N #}N 'writing instrument'
12. { [ cey ] V # [ inai ]N # } N 'sorcery'
13. { [ varai] V # [ paTam] N # }N 'map, plan'
14. { [ teey ]V # [ piRai ] N # }N 'waning moon'
15. { [vaLar] V # [ piRai] V # }N 'waxing moon'
16. { [ periya ]adj # [ nakaram ]N # }N 'big city'
17. {{ nalla ]adj [ puttakam ]N #} N 'good book'
18. { [ veLLai ] adj [ uLLam] N #} N 'innocent heart/man'
19. { [ inRu] adv # [ kaalai] N # } N 'today morning'
20. { [ naaLai] adv # [ maalai] N # }N 'tomorrow evening'
21. { [ neeRRu] adv # [ iravu] N #} N 'yesterday night'

It has been stated that the "IS A" condition derives the category which is the same as that of 'C' if there are two constituents like 'a' and 'C'. According to this condition [eLhtu kool] is an instrument; [kalvi kuuTam] is an institution. But there are a few compounds where the "IS A" condition derives the category as that of the constituent 'C' but semantically it may mean something different from the meanings of the constituents as shown in the following examples:

22. { [ kampi] N # [ eNNu] V # }V 'to undergo imprisonment' (literally, rod count, or count the rod!)
23. { [ kampi] N # [ niiTTu] V # }V 'to decamp' (literally, rod lengthen, or lengthen the rod!)
24. { [ kaatu] N # [ kuttu] V # }V 'to tell fibs' (literally, ear pierce, or pierce the ear!)
25. { [ kuLir] N # [ viTu ]V # }V 'to outgrow fear' (literally, chill leave, or leave the chill!)
26. { [ kai] N # [ nanai ]V # }V 'to dine at one's house' (literally, hand wet, or wet the hand!)

It has been shown here that the Dictionary of Tamil will have stems and words. Word Formation Rules operate on the stems and words to derive simple words and compounds. Derivational Rules and Compounding Rules are not part of the Word Formation Rules. So, the model of the proposed Dictionary5 of Tamil will look the shown below.

Lexical Component

↓

Dictionary

↓

Words Stems

↓

Word Formation Rules

↓

Derivation Rules

↓

Compounding Rules

↓

Inflectional Rules

↓

Readjustment Rules

↓

Truncation Rules

↓

Allomorphy Rules

↓

Output

NOTES

1. Three formatives have been proposed here, namely, morphemes (suffixes) stems and words, but there are only two boundary markers, namely, morpheme boundary ( + ), and word boundary ( # ). It is not clear whether internal word boundary ( # ), to add stems with words is needed as in mara #, niLhal 'shade of tree,' mana #, kooTTai 'dream of future.' This has to be worked out further and decided.

2. Having stems as separate formatives from words may avoid too many phonological rules in between WFR's.

3. Though the paper is restricted to Derivation and Compounding, past tense markeers are used to state the phonological restrictions considering this fact as a general information for Tamil morphology.

4. X - may represent any formative: morpheme or stem or word.

5. Though all types of WFR's are not explained here, the schematic representation gives a rough idea about all types of WFR's of the lexical component.

BIBLIOGRAPHY

1. Aronoff, Mark. 1976. Word Formation in Generative Grammar. MIT Press, Cambridge, Mass.

2. Booij, G. E. 1977. Dutch Morphology: A Study of Word Formation in Generative Grammar. The Peter De Ridder Press, Lisse.

3. Kageyama, Taro. Word Formation in Japanese. Lingua 57, Nos. 2-4. pp. 215-258.

4. Kiriyaavin tarkaalattamiLh akaraaati. 1992. Cre-A, Chennai.

5. Lehman, Thomas. 1989. A Grammar of Modern Tamil. Pondicherry Institute of Linguistics and Culture, Pondicherry.

6. Ramaswami, N. Derivation and Inflection in Bhumij. (Manuscript). Central Institute of Indian Languages, Mysore.

7. Ramaswami, N. Diglossia: Formal and Informal Tamil. (Manuscript). Central Institute of Indian Languages, Mysore.

8. Scalise, Sergio. 1984. Generative Morphology. Foris Publications, Dordrecht, Holland.

9. Spencer, Andrew. 1991. Morphological Theory. Basil Blackwell, Cambridge.

10. Subrahmanyam, P. S. 1971. Dravidian Verb Morphology: A comparative study. Annamalai University, Annamalainagar.

*** *** ***

HOME PAGE | Multilingualism in India | Teaching Indian Languages Through Correspondence | Look Into My Eyes: What Do You See? | A Review of the Felicitation Volume in Honor of Professor Karunakaran | CONTACT EDITOR

N. Ramaswami, Ph.D.
Director i/c (Retired)
Central Institute of Indian Languages
Mysore 570006, India